Route conversations between bot and human agents

Botstar chatbots hand a conversation to a live agent when a rule triggers it, such as a customer asking for a refund or a message containing a specific keyword. A Beam agent watches the handoff queue, reads the conversation history and contact details Botstar has stored, and applies the approved routing rule to assign the conversation to the right team or queue. It can also log the outcome once the conversation closes. Anything outside the defined triggers, such as an ambiguous request or a first-time keyword, gets routed to a human to decide rather than handled automatically.

Update bot flows and stored responses

Botstar stores each bot as a flow of steps, messages, and conditions across channels like a website widget or Messenger. A Beam agent can read the current flow configuration, apply an approved change such as updating a saved reply, retiring an old prompt, or turning a step on or off, and write that update back. Triggers might include a scheduled review, a product change, or a spike in a specific customer question. Any change to logic that affects how the bot branches or escalates is left for a person to approve before it goes live, since flow structure is harder to verify from data alone.

Read conversation and contact data across channels

Every chat handled by Botstar creates a conversation record tied to a contact profile, channel, and status. A Beam agent reads this data on a schedule or when a conversation closes, checking for patterns such as repeated unresolved topics or a contact who has messaged multiple times without a reply. It can update contact records, tag conversations, or notify a channel when a rule is met. Where a conversation involves a complaint, a request outside approved categories, or unclear intent, the agent leaves the case in the queue for a support agent to answer directly.